Smart Monitoring for Long-Term Performance of Green Stormwater Infrastructure

Linking smartphone images, hydrologic screening, and maintenance decisions.

Three-step smart monitoring workflow showing photo capture, AI screening, and maintenance prioritization
The project workflow links field photos and notes to performance screening and maintenance priorities. Field photos were taken by the PI, and the cartoon figures were generated using AI.

This project will develop and test an AI-enabled, smartphone-based monitoring workflow for green stormwater infrastructure. Students, residents, community groups, and stormwater managers will be able to take standardized site photos and field notes that help identify maintenance needs.

The workflow will use vision-language models to screen for visible concerns such as blocked inlets, sediment buildup, unhealthy vegetation, litter, standing water, and erosion. Surveys in Boston and New York City, together with focused hydrologic checks at selected Boston sites, will connect visual conditions with drainage performance. The goal is to detect maintenance needs earlier and help rain gardens, bioswales, and related systems function effectively over time.

This project is conducted in collaboration with and supported by the Stone Living Lab (SLL).